Aims: The risk of incident atrial fibrillation (AF) can be estimated by clinical parameters in the Framingham AF risk model. Elevated N-terminal pro-B-type natriuretic peptide (NT-proBNP) and increased rate of premature atrial contractions (PACs) have been shown to be associated with AF, but the additive value of both of these biomarkers in the Framingham AF risk model has not been fully examined.

Methods and results: A total of 646 subjects from the Copenhagen Holter Study (mean age 64.4 ± 6.8 years, 41.6% women) with no history of prior AF, stroke or cardiovascular disease were followed for the diagnosis of incident AF or death (median follow-up time 14.4 years). Median NT-proBNP was 6.7 pmol/L (IQR: 3.6-13.5), median PAC count was 1.4 beats/h (IQR: 0.6-4.5), 71 (11.0%) subjects developed AF, and 244 (37.8%) died. Multiple Cox regression including Framingham AF risk score, log-transformed NT-proBNP, and log-transformed PAC showed a significant increase in AF hazard risk [hazard ratio (HR) 1.45, 95% confidence interval (CI) 1.14-1.85, P = 0.002; HR 1.23, 95% CI 1.09-1.39, P = 0.001]. The addition of PAC to the Framingham AF risk model significantly improved the time-dependent area under the receiver operating characteristic curve (AUC 65.6 vs. 72.6; P = 0.008), while the addition of NT-proBNP did not.

Conclusion: Atrial fibrillation risk discrimination was significantly improved by the addition of PAC to the Framingham AF risk model, but not by the addition of NT-proBNP.